Spi is essential for both enhancer-mediated and basal activation of ...

Nucleic Acids Research, 1994, Vol. 22, No. 4 669-677

Spi is essential for both enhancer-mediated and basalactivation of the TATA-less human adenosine deaminasepromoter

Mary R.Dusing and Dan A.Wiginton*Department of Pediatrics, Division of Basic Science Research, University of Cincinnati College ofMedicine and Children's Hospital Research Foundation, Cincinnati, OH 45229, USA

Received August 25, 1993; Revised and Accepted January 14, 1994

ABSTRACT

Tissue-specific expression of the human adenosinedeaminase (ADA) gene Is mediated by transcriptlonalactivation over a thousand-fold range. C/s-regulatoryregions responsible for high and basal levels ofactivation Include an enhancer and the proximalpromoter region. While analyses of the T-cell specificenhancer have been carried out, detailed studies of thethe promoter region or promoter-enhancer Inter-actions have not. Examination of the promoter regionby homology searches revealed six putative Sp1binding sites. DNase I footprint!ng showed that Sp1 Isable to bind these sites. Deletion analysis indicated thatthe proximal Sp1 site Is required for activation of areporter gene to detectable levels and that the moredistal Sp1 sites further activate the level of expression.Inclusion of an ADA enhancer-containing fragment inthese deletion constructions demonstrated that Sp1sites are also essential for enhancer function. Appar-ently Sp1 controls not only low level expression but isalso an integral part of the mechanism by which theenhancer achieves high level ADA expression. Muta-genesis of a potential TBP binding site at base pairs- 21 to - 26 decreased activity only two-fold indicatingthat It is not essential for transcriptlonal activation orenhancement.

INTRODUCTION

Adenosine deaminase (ADA) is an enzyme in the purinemetabolic pathway which catalyzes the irreversible deaminationof adenosine and deoxyadenosine to inosine and deoxyinsinerespectively. It is expressed in all human tissues in a definedpattern over a thousand fold range (1—4). Among these tissuesthe highest levels are observed in thymus where it plays a criticalrole in T-cell survival by scavenging excess deoxyadenosine thatis toxic to these cells. Its absence results in a form of severecombined immunodeficiency (SCID), a rare genetic disordercharacterized by lack of T-cell function and impaired immunefunction (5,6). Unlike genes expressed in one tissue, or at one

developmental time, which are regulated in an on—off fashionby binding of tissue or stage specific trans-acting factors, the ADAgene must accomplish differing levels of expression in multipletissues as development and differentiation occur. Thus it providesan interesting model for a gene which is expressed in all cells,yet still can achieve activation to very high levels in some andexpression at lower but highly modulated levels in others.

Regulation of ADA levels in human cells occurs mainly at thelevel of gene transcription (7). Transcriptional initiationpresumably serves as the key point of regulation, althoughtranscriptional arrest or pausing has been proposed as having asignificant role in some cell types (8). The identification andcharacterization of the cis-regulatory elements and the trans-actingfactors that interact to achieve appropriate spatial and temporalexpression of the human ADA gene have been the recent focusof our research. Two functional DNA segments, a complexregion from the first intron with T-cell specific enhancer activityand a proximal region upstream of exon one with promoterfunction, have been identified thus far. Characterization of the~ 200 bp enhancer region revealed that a number of proteinsfunctionally bind this segment (9). Analysis of these enhancercomponents is underway. The characterization of the humanADA promoter segment has been far less detailed. A 232 bppromoter segment from the 5' end of the ADA gene has beenshown to be effective in activating a chloramphenicol acetyltransferase (CAT) reporter gene to low basal levels in transientassay (4,10). High levels of CAT expression were obtained withthis promoter in T-cells, in both transient assay and transgenicmice, in the presence of the ADA enhancer (4,9). However invivo, in transgenic mice, the facilitator segments flanking eitherside of the enhancer are required for full, copy-numberdependent, insertion-site independent expression of the transgenein thymus (9). The facilitators seem to assist in formation of theactive chromatin region associated with the enhancer. A recentreport indicates that the /3-globin locus control region(LCR)/enhancer requires a functional promoter segment intransgenic constructions to consistently produce the chromatinactivation associated with the LCR (11). Since the promoter ofthe human ADA gene has previously been only superficially

*To whom correspondence should be addressed

Downloaded from https://academic.oup.com/nar/article-abstract/22/4/669/1194725by gueston 14 April 2018

670 Nucleic Acids Research, 1994, Vol. 22, No. 4

characterized, we undertook a detailed delineation of itscomponent parts. In this way we hoped to gain insight not onlyinto basal promoter function, but also into what is required ofthe promoter elements for promoter/enhancer interactions andT-cell enhancer function.

Sequence analysis has previously identified six putative bindingsites for the transcription factor Spl in the proximal region ofthe human ADA promoter (10,12). Similar Spl binding sites havebeen found in a wide variety of mammalian and viral promoters(13 and references therein). The functional nature of the putativeSpl sites in the human ADA promoter has not been investigated.The examination of the human ADA promoter sequence alsorevealed a sequence similar to the consensus TATA bindingprotein (TBP) binding site that is situated in the proper locationto bind that component of the basal transcriptional machinery(14). Our studies were designed to elucidate whether theseelements or others were required for basal activation of the ADApromoter. Dissection of the promoter into its functional partsenabled us to determine which element(s) are required, whichare supplementary, and which are unnecessary for basalactivation. With this accomplished, we then investigated therelationship between other defined regulatory regions of the gene(e.g. the enhancer) and specific promoter elements which arerequired for enhanced activation of transcription. With thisinformation, we can begin to understand the hierarchy of ADAregulatory components and start to design a basic model for thecomplex regulation of this gene.

MATERIALS AND METHODSPreparation of ADACAT expression constructions

The previously described plasmid p AD AC AT 2.2 (4) was cutwith Ndel, blunt-ended with Klenow and cut with BamHI toisolate the 3.8 kb fragment containing a 2.2 kb segment of thehuman ADA gene fused to the CAT coding sequence. Thisfragment was ligated into HindlH-cut, Klenow-blunted, BamHI-cut pUC18 plasmid DNA. The resulting plasmid was digestedwith Hindin and BssHII and the 4.6 kb fragment was blunt-endedwith Klenow and recircularized to generate pl8ADACAT. Thisplasmid contains an ADA promoter fragment from —2094 to+ 97 cloned directly upstream of the CAT reporter gene. Allother plasmids described are derived from it.

5' truncations of the ADA promoter were generated bydigesting pl8ADACAT with BssHII and Bal 31 for variousamounts of time. DNA was blunted with Klenow and dNTP's,Hindm-linkered and digested with BamHI and Hind in. Thetruncated 1.9 kb fragment was isolated and ligated to the 2.6 kbHindm-BamHI fragment of pl8ADACAT. The clonesgenerated were sequenced to identify deletion break points.Selected clones were chosen such that putative promoter elementswere partially or completely deleted.

Internal promoter deletions were generated in the followingmanner. A 4.58 kb Apal fragment from pl8ADACAT wasrecircularized to form A(71 — 34) with an internal deletion of Splsites I, II and HI. Larger internal deletions were generated fromA(71—34) by digesting with Apal and Bal31 for very limitedlengths of time. DNA was blunt-ended with Klenow and digestedwith BamHI. The resulting fragments were both isolated and eachwas ligated to the corresponding Apal (blunted)-BamHIfragment from A(71 —34) such that unidirectional deletions fromthe Apal site of A(71—34) were obtained in both directions.A(96-34) removes Spl sites IV and V, A(108-34) removes

all of the remaining Spl sites, and A(71 -10) removes the TBPbinding site.

To create a specific mutation of the potential TBP site, aplasmid with an internal promoter deletion of 65 bp (Fig. 1,residues —74 to —10) was utilized. In this plasmid (not shownin Figure 5) a Not I site is serendipitously created at the deletionjunction. The plasmid was digested with Not I and mung beannuclease. A synthetic double-stranded oligonucleotideGGGCCCGGCCCGTCTGCAGGAGCGTGGCCGGCC wasphosphorylated, ligated onto the blunt ends of the fragmentdescribed above and digested with Pstl. A 4.6 kb fragment wasisolated and self-ligated. Resulting plasmids were cloned andsequenced. A clone was isolated which had the oligonucleotidein the desired orientation. This plasmid was cut with Apal andBssHII. The 4.5 kb fragment was isolated and ligated to a 172bp Apal-partial BssHII fragment from pl8ADACAT to makepl8ADACAT-T. This plasmid contains exactly the samesequence as pl8ADACAT except that the TBP site TAAGAAhas been replaced by a Pstl site, CTGCAG, which has nohomology to the TBP consensus binding sequence.

This original series of deletions was formed in a pUC18plasmid backbone and gave questionable results upon transfectioninto cultured cells. To remedy this we chose to use the plasmidpBLCAT6 obtained from G.Schiitz (15). Plasmids with thepromoter truncations, deletions, or mutations described abovewere digested with either Hindin (5' truncations) or BssHII(internal deletions and TBP mutation), blunt-ended with T4polymerase and cut with Ncol. The 900-600 bp fragments ofthese digest were ligated to 3.7 kb Xhol (blunted) -Ncol fragmentfrom pBLCAT6 recreating the series of deletions/mutationsdescribed above in the pBLCAT6 backbone. These clones weredesignated pADACAT211 and derivatives and are shown in Figs.5 and 6.

An enhancer fragment was isolated from pSph2.3-, a pUC18plasmid containing a 2.3 kb SphI fragment of the human ADAfirst intron enhancer (GenBank #8327-10584) cloned into theSphI site of pUC18. This 2.3 kb fragment was removed bydigestion with Sail and Hindin. Fragment ends were repairedwith T4 polymerase and dNTP's, Clal-linkered and subclonedinto selected pADACAT211 promoter deletion/mutation plasmidsat the Clal site 3' of the CAT reporter gene. This generatedplasmids containing various promoter segments and the T-cellspecific enhancer fragment downstream of CAT in its nativeorientaion (See Fig. 7). The enhancer fragment described wasalso cloned into pBLCAT6 to create a promotorless enhancerconstruction.

Plasmids were grown in DH5 a E.coli (Bethesda ResearchLaboratories) and purified by banding twice in CsCl gradients.All plasmids were sequenced by double-stranded dideoxy methodswith Sequenase (U.S.Biochemical).

Transfection and CAT assays

A modified DEAE-dextran transfection protocol describedpreviously (4) was used to introduce deleted plasmids into MOLT4, an immature human T-cell line with high ADA levels (1650nmoles/min/mg protein) or Raji, a human B-cell line with lowADA levels (14 nmoles/min/mg protein). Cell culture and CATassays were done as previously described (4,9). All samples weretransfected in duplicate. The resulting CAT activities for duplicatesamples differed by less than 10%. The average value of theduplicates was used for the calculation of results. The absoluteCAT activity for a particular construction, preparation of DNA,


Nucleic Acids Research, 1994, Vol. 22, No. 4 671

or cell line transfected varied between experiments, probably asa result of transfection efficiency and cell state. However, therelative changes in CAT activity compared to controls were nearlyidentical.

Extract preparationThe plasmid, pSpl-516C, a generous gift from R.Tjian, wastransformed into XL 1-Blue E.coli (Stratagene) and bacterialextracts containing recombinant Spl were isolated as in Kadonaga(16). XL 1-Blue extracts without the plasmid were made by thesame method to serve as control. MOLT 4 nuclear extracts wereprepared as described previously (9).

Electrophoretic mobility shift assay (EMSA)A radiolabelled 128 bp Eco52 I-EcoRI fragment containing thehuman ADA promoter was prepared as for footprinting. 200 pgof this fragment was mixed with 3.6 /tg of MOLT 4 nuclearextract in a volume of 25 /il at a final buffer concentration of18% glycerol, 70 mM KC1, 33 mM Tris-HCl pH 7.9, 10.5mM MgCl2, 0.9 mM EDTA, 0.9 mM DTT, 0.5 mM ZnCl2,0.4 mM sodium metabisulfite, 80 /*M PMSF, and 2 /tg poly (dl-dC)-poly (dl-dC). 30 ng of a double-stranded Spl consensusoligonucleotide ATTCGATCGGGGCGGGGCGAGC or an Splmutant oligonucleotide ATTCGATCGGTTCGGGGCGAGC(Santa Cruz Biotechnology) were added to the EMSA reactionwith the labelled probe as competitors. For antibody supershiftanalysis, 3.6 /ig of MOLT 4 extract was incubated overnight at4°C with 5 jig of rabbit polyclonal IgG to either Ets-l/Ets-2 orSpl (Santa Cruz Biotechnology). Extract—antibody mixtureswere then used for EMSA. Complexes were separated byelectrophoresis through 3.8% non-denaturing polyacrylamide gelsat 4°C, 20 mA in 25 mM Tris base, 190 mM glycine, 1 mMEDTA.

DNase I footprintingFragments used for footprinting were isolated from pADA-CAT211. Either a 233 bp Eco52 I-BamHI fragment or a 128bp Eco52 I-EcoRl fragment was used to generate the upperstrand footprint. The lower strand was footprinted with a 270bp BamHI-Bsal fragment. Klenow, [a-32P] dGTP(3000Ci/mmole; NEN) and dNTP's were used to fill in the Eco52I and BamHI site respectively after digestion of pADACAT211with these enzymes. Plasmids were then cut with the secondenzyme and the fragment of interest was isolated from low meltagarose by melting, phenol extraction, and precipitation. DNaseI footprinting was performed as described previously (9) using0-20 jig of crude bactrial extract containing rSpl or 0—18 /tgof MOLT 4 crude nuclear extract.

RESULTSAnalysis of ADA promoterA 232 bp segment from the 5' end of the human ADA gene hasbeen shown previously to be effective at activating a CAT reportergene in transient assay in a number of different cell types (4,10).In transgenic mouse systems this promoter segment is also capableof driving ubiquitous tissue expression of the CAT reporter, aswell as high level thymic CAT expression (9). This pattern oftransgene expression was only observed from constructions thatalso contain a fragment encompassing the ADA first intronenhancer region. A segment of this promoter is shown in Fig.1. Like many genes with GC rich promoters, the ADA gene has

multiple transcriptional start sites. One major and several minorstart sites have been identified in both mouse (17 — 19) and human(10) by several different techniques. Sequence analysis of thisDNA fragment from the human gene revealed 6 potential Sp 1binding sites (10,12). These sites, labelled I -VI , are shown inFig. 1. Also shown is the location of a potential TBP bindingsite, TAAGAA, and the major transcriptional start site.

An homology comparison between the murine and humanADA promoter regions has been published previously (20).Limited regions with significant homology were identified, mostlydue to the GC rich nature of both promoters. Unlike the proximalpromoter of many other genes, this area does not seem to behighly conserved between the species. The functional elementsappear similar but not identical. The murine gene has five non-overlapping Spl sites whose sequence and position relative tothe major transcriptional start site differ significantly from thesix overlapping Spl binding sites in the human promoter. Themurine gene also has a recognizeable TATA motif, TAAAAAA,which has been shown to bind to the TFHD fraction (21). Thehuman gene sequence at that location, TTAAGAA, is moredivergent from the canonical TBP binding sequence (22). It hasbeen questioned whether or not the TAAGAA sequence mightfunction in TBP binding (12). In fact, the human adenosinedeaminase gene is often cited as containing a TATA-less promoter(10,13,23,24).

Electrophoretic mobility shift assay (EMSA)The ability of a small 128 bp fragment of the human ADApromoter to bind protein(s) in MOLT 4 nuclear extracts wasassessed by EMSA and is shown in Figure 2. Formation of theshifted complexes designated by small arrowheads (Fig. 2, Lane2) was inhibited with an excess of an unlabelled oligonucleotidecontaining an Spl consensus binding site (Fig. 2, Lane 5) butnot a similar amount of an oligonucleotide containing a mutationin the Spl site (Fig. 2, Lane 6). Addition of a polyclonal antibodyagainst Spl resulted in a reduction of the intensity of the of theSpl specific bands and the concomitant appearance of asupershifted band (shown by the large arrowhead in Fig. 2, Lane3) indicating the presence of Spl in these complexes. The addition

5 ' OCOCOCCCACAOOTOOOTCCOOTCOOOCOCCOCOOWJCI I

- 2 1 0 - 1 8 0

COTAOTTTTCOOOTCO ICOAOOACOCCOOOTCCAOAATTCCA

-150V

LTOCOCOATCCM

- 1 2 0

rv

CCOOCOA

- 9 0

OC0- 6 0

CCOQCCCQTK'iuuuuuiAOcaTooccaaccocaaccAccocTooccc 3 '-30

Figure 1. Human ADA promoter sequence. The sequence proximal to the ADAmajor transcription start site (+1), indicated with the bent arrow, is shown from-211 to +ll.Thisccinespcindstothesegiiiemfinom3725to3^46intheGenBanknumbering for the human ADA gene. This promoter lacks a canonical CCATbox sequence. The sequence from -21 to —26, indicated with a broken line,has low homology to the TATA binding protein (TBP) consensus binding siteand is in the proper location to bind TBP. Six consensus matches for the trans-acting factor Spl are shown beneath the solid lines and are labelled I—VI.



antibody

competitor

extract

urn- uti-Spl Ell

Spl mSpl

*" •*I *

Figure 2. Electrophoretic mobility shift assay (EMSA) of the human ADApromoter with MOLT 4 extract. Labelled probe was made as described. 200 pgwas incubated alone (Lane 1) or with 3.6 /ig of MOLT 4 nuclear extract (Lane2). Binding competition was performed with an 800-fold molar excess of anoligonucleotide containing one consensus Spl binding site (Lane 6) or an identicalamount of an oligonucleotide containing a mutated Spl site (Lane 7). 3.6 fig ofMOLT 4 extract was also incubated overnight at 4°C with either 5 jig of polyclonalrabbit antj-Ets-l/Ets-2 igG (Lane 4) or 5 /jg of polyclonal rabbit anti-Spl IgG(Lane 3). Labelled probe was then added and allowed to incubate at 20°C.Complexes were separated by electrophoresis through non-denaturingpolyacrylamide gel at 4°C. Small arrowheads denote shifted complexes whichcan be competed with an Spl oligonucleotide but not the mutant. Antibodysupershifted complex is indicated with the large arrowhead,

of a polyclonal antibody to the Ets family of transcription factorsresulted in no bands with reduced intensity or supershifted bands(Fig. 2, Lane 4).

Footprinting analysisThe sequence and position of the consensus Spl binding siteswithin the human ADA promoter are shown in Fig 1. All sitespossess the GGGCGG core binding sequence characteristic ofGC boxes with the exception of site I which has the sequenceAGGCGG. Homology to the Spl consensus binding site,[(G/T)(G/A)GGCG(G/T)(G/A)(G/A)(C/T)] (25), varies from80—100%. Elements with sequence identical to sites I and Vin the ADA gene promoter have been shown to bind Spl in thecontext of other promoters where they were found (23,26,27).The ability of the remaining sites to bind Spl was unknown. Inorder to assess this ability and to determine relative affinities ofthese sites for Spl, DNase I footprinting was performed. Arecombinant Spl -lac Z fusion protein from pSpl-516C (16)containing the Spl DNA binding and activation domains wasproduced in bacterial cultures and extracts were used to footprintradiolabelled ADA promoter fragments containing all six putativeSpl binding sites. All potential Spl sites are protected fromDNase I digestion on both strands by pSpl-516C extract (Fig.3). XL 1-Blue bacterial extract without Spl did not footprint thisregion (data not shown).

The footprinted region produced with MOLT 4 nuclear extract(shown in Fig. 4) varies from that generated with bacterial extract

B

••

II

•3!

VI & V rv

OCOOOCOOOOCOOOAOOCOOOOCCCOOCCCOTTOCCCOTTAAOAJU3AOCOTO(X^S»XX^XX30CCACCOCTOOCCCOOCOCCOOTOOCOACCOOa

III -I

Figure 3. DNase I footprinting of the human ADA promoter with recombinantSpl (rSpl). Fragments for footprinting were generated from pADACAT211 andlabelled as described in the methods. Bacterial extracts containing rSpl wereincubated with end-labelled fragment and limited DNase I digestion was performed.The resulting fragments were separated on 6% sequencing gels. Locations ofthe Spl sites I-VI are marked with thin solid lines. Protected regions are indicatedwith heavy lines. (A) A 233 bp Eco52I—BamHI fragment was used for generatingthe coding strand footprint. Lane 1 is a Maxam - Gilbert C+T sequencing ladderof the fragment. Lane 2 contains no bacterial extract and lanes 3 - 5 contain 5^g,10/ig, and 15/tg, respectively, of bacterial extract containing rSpl. (B) A 270bp BamHI-Bsal fragment was used to footprint the non-coding strand. Lane1 contains the C+T ladder for the fragment and lanes 2 - 6 contained Ojig, 5/ig,10/ig, 15(ig, and 20/ig of bacterial extract. (C) Protected areas and consensusSpl sites are indicated relative to the promoter sequence. Increased sensitivityof a site to DNase I digestion is marked with an asterisk *.

containing recombinant Spl. At intermediate levels of extract,the entire proximal promoter region from site IV to beyond thetranscription start site becomes DNase I resistant (Fig. 4, Lane7). This may be due to additional proteins that are able to interacteither with this region of the promoter or with the bound Splproteins. The most notable difference with the crude nuclearextract is the absence of a protected region over sites V and VIeven at the highest protein levels examined. This may be relevantto the results of the deletion analysis described below.

Promoter deletion analysisIn order to assess the functionality of these Spl sites in a transientassay system, the plasmid pADACAT211 was made, along withvarious 5' truncations and internal deletions of the promotersegment affecting the Spl sites and the potential TBP site (Fig.5). These plasmids were tested for their ability to activate theCAT reporter gene after DEAE-dextran transfection into thehuman lymphoid cell lines MOLT 4 and Raji. These cell lineshave both been shown to efficiently utilize the intact human ADA



Py G

VI

IV

MI

Rrfrtv*CAT Activity

KAJ HOLT4

Figure 4. DNase I footprinting of the ADA promoter with MOLT 4 extracts.Footprindng was performed as described for Fig. 3 using MOLT 4 nuclear extracts.Spl sites are indicated as in Fig. 3. Maxarn-Gilbert C+T and G sequencingladders are shown in Lanes 1 and 2. Lanes 3 - 9 contain 0 /ig, 0.9 y.g, 1.8 /ig,4.5 fig, 9 /ig, 12.6 /ig, and 18 ng respectively.

promoter in this transfection-transient assay system (4). Cellextracts were prepared 48 hours after transfection and wereassayed for CAT activity. The original series of deletions wasformed in a pUC18 plasmid backbone and gave anomalous resultsupon transfection into cultured cells. Removal of promoterelements was not sufficient to destroy CAT expression. Thisseemed to be due to cryptic transcriptional starts within theplasmid which severely interfered with interpretation of results.Previously in our lab, similar results (unpublished data J.C.States,D.A.Wiginton, and J.J.Hutton) had been seen using the pSVO-CAT vector backbone (28). To remedy this we chose to use theplasmid pBLCAT6 (15), in which the plasmid sequencesresponsible for cryptic starts have been deleted and replaced withtwo polyadenylation signals upstream of the 5' multiple cloningsite. This efffectively prevents upstream starts, as evidenced bythe fact that the promoterless parent vector had no activity inany cell lines tested. DNA segments from the parental plasmidand all subsequent deletions were sublcloned into pBLCAT6 asdescribed in the methods to generate pADACAT211 and thepADACAT deletions shown in Figs. 5 and 6.

Relative CAT activities obtained after transfection of theseplasmids are also shown in Figs. 5 and 6. Values for CAT havebeen normalized to the parental construction (top in each figure).The deletion of sites V and VI (Fig. 5, A100 and A95) has littleeffect on resulting CAT activity. This correlates well with theabsence of a DNase I protection footprint over these sites usingMOLT 4 extracts and suggests that these sites may not bind Splin vivo even though they have the capacity to do so. Deletionof site IV (Fig. 5, A75) reduces activity 6 - 7 fold in both MOLT4 and Raji. Deletion of site HI (Fig. 5, A52) reduces activityanother 3 - 4 fold. Converting site III into a better consensus Splsequence, GGGGCGGGGC (Fig. 5, A56) by replacing the first

pADACATZl!

Aid

A14I

4100

4»5

asi

475

AM

ajj

040

AJ4

414

A4

W71-M)

————mi- •

100

1Z3

U

1S1

110

111

19

za

4

„,<1

<1

100

t*

111

157

195

174

24

33

7

•a

<2

<2

A(10*-34)

&T71-10)

Figure 5. CAT activity of transfected promoter deletions. The structure of theproximal promoter region of Bal 31 -deleted constructions from the pADACAT211series are depicted graphically in Fig. 5. The constructions with 5'-truncationsof the promoter are designated with a An where 'n' is the length of the remainingpromoter segment. The constructions with internal deletions are designated byindicating the negative numbers corresponding to the residues immediately adjacentto the deleted segment. Spl sites are indicated by filled boxes, the TBP bindingsite by a hashed box, the major transcription start site with the bent arrow, andinternal deletions with a broken line. Spl sites shown represent those indicatedas sites I -VI (right to left) in the text. These constructions were transfected intoMOLT 4 or Raji cells. Cell extracts were prepared and assayed for CAT activity.The activities were normalized to that obtained for the parental construction,pADACAT211 (at the top of Fig. 5), which was set at 100. In clone A56 thefirst C of Spl site Hi was deleted and the juxtaposed G (*) results in an improvedconsensus in the Spl site of GGGGCGGGGC.

residue (C to G) increases activity slightly relative to A75. SitesI and II were not separated by any of the deletion clones weobtained, but their simultaneous deletion (Fig. 5, A40 and A34)results in loss of detectable CAT activity in either cell line. Theseresults would indicate that sites I and/or II alone are capable oflow level basal activation of the promoter, and addition of sitesIE and IV further activates transcription to its maximalunenhanced level. The addition of sites V and VI do notappreciably increase activation above this level in our assaysystem.

Interpretation of the results of the internal deletions is lessstraightforward than the truncations. Each of these constructionsis inherently different from the parental construction in theplacement of the remaining Spl elements relative to thetranscriptional start site. Internal deletions A(71—34) andA(96-34) maintain the spacing between the transcriptional startsite and the most proximal remaining Spl site at a distance nearthat found in the endogenous ADA gene (—36 bp), although thesequence for the Spl site found in this position varies. An internaldeletion removing sites I—HI [Fig. 5, A(71 -34)] placed site IVat - 3 5 bp, a distance almost identical to the location originally



MatinCAT Activity

PA0ACAT211

PMMCAT211T

fl(7l-)4)T

/K71-34)

O-CTOCAt-

Figure 6. CAT activity of transfected TBP mutants. Devices used to depictpromoter elements are as in Fig. 5, with the addition of an open box representingthe mutant sequence, CTGCAG, whose presence is designated by the additionof T to the name of the construction. This has been substituted in some clonesfor the native sequence TAAGAA, depicted with the hashed box. Resulting CATactivity is shown normalized to pADACAT211 which was assigned a relativevalue of 100.

occupied by site I. This construction was unable to activatetranscription to a detectable level in either cell line even thoughit has three Spl sites (the same number as A75), the closest ofwhich is the same relative distance from the transcriptional startsite as sites I/TJ in the parental construction. This implies thatthese Spl sites, IV—VI, are functionally different in their abilityto substitute for the more proximal sites, I—in, and activatetranscription. A larger internal deletion positioning site VI alone[Fig. 5, A(96—34)] at - 3 5 bp results in very low level activationat best. Deletion of the putative TBP binding site [Fig. 5,A(71-110)] moves sites IV-VI to - 1 1 bp, -31 bp, and - 3 6bp respectively. This construction also produces very low CATactivity. The overall results also demonstrate the absence of othermore distal sequences which are necessary or sufficient foractivating transcription.

Mutation of the potential TBP siteThe sequence TAAGAA is found in the ADA promoter at —21to — 26 bp relative to the major transcription start site at the properlocation to function in the binding of TBP. A similar sequence,TAAAAAA, is present in the murine ADA gene at - 21 to - 2 7bp and has been reported to be essential for promoter function(21). In order to determine the necessity of a similar sequenceat that site for activation of the human ADA gene, it was mutatedby clonal manipulation from TAAGAA into a PstI site, CTGC-AG [Fig 6, pADACAT211T and A(71-34)T] and tested intransient assay by transfection into MOLT 4 and Raji cells. Thismutation alone did not obliterate activity, although a smalldecrease in activation of 50% or less was observed in both Rajiand MOLT 4. While modification of this sequence has adiscernible affect, this specific site is not essential for promoterfunction.

Promoter requirements for enhancer function in MOLT 4The ADA T-cell enhancer was included in selected plasmids(shown in Fig. 7) downstream of the CAT coding sequence inits natural orientation relative to the promoter. The enhancer hasbeen shown previously to be utilized very efficiently in MOLT4 but not Raji (4). Therefore, promoter-enhancer constructionswere transfected only into MOLT 4 cells to examine whichpromoter elements, if any, are required for enhancer drivenactivation. The plasmid pBLCAT6enh, which contains theenhancer but lacks the ADA promoter, showed no detectable

100

54

<1

100

87

<2

<2

pAOACATZUartfl

A75«Vi

pADACAT 211T«*I

fUhtiv*CATAttMty-

HM.T4* «•

100 (2200)

4S (10(0)

1 (Z1)

0.4 (t)

2 (41)

40 (870)

• • ipl D--CT0CA8-

Figure 7. Structure and relative CAT activity of promoter-enhancer constructions.The promoter region for these constructions is as shown above. These promoterregions are identical to some of those shown in Figs. 5 and 6. These plasmidsalso contain, downstream of the CAT sequence, a 2.3 fcb ADA intronk fragmentthat contains a T-cell specific enhancer (indicated by the addition of enh to thenames). Transfectkms were done with MOLT 4 cells only and the CAT activitiesare reported two different ways: normalized to either pADACAT21 lenh (*column)or pADACAT211 (•• column).

activity (data not shown). Inclusion of the ADA transcription start(Fig. 7, A14enh) gave activity at very low but detectable levels.Addition of the 'TBP' site (Fig. 7, A40enh) increased activityan additional two-fold. The level of activation seen in the presenceof the enhancer and Spl sites reflects a further 50—100 foldincrease in CAT activity over A40enh (Fig. 7, pADACAT21 lenhand A75enh). All the constructions containing the enhancer showa 20 to 50 fold increase above the equivalent plasmids lackingthe enhancer. Results of this experiment show that the enhanceris capable of recognizing and activating the truncated promoter(Fig. 7, A40enh and A14enh) to low levels even in the absenceof Spl sites. However, higher levels of activation require thepresence of both the enhancer and the Spl binding sites.

The TBP mutant pADACAT21 ITenh shows a 60% decreasein activity from pADACAT211enh. This is very similar to theresults seen in the absence of the enhancer, demonstrating thatneither the Spl nor the enhancer activated transcription has anabsolute requirement for this sequence.

DISCUSSION

Much of the regulation of human ADA gene expression occursat the transcriptional level (7). In addition to regulation oftranscriptional initiation, transcriptional pausing and arrest havebeen observed in the human and mouse ADA genes in tissuesand cells with low levels of endogenous ADA (7,8,29). Whilethe sequences responsible for the observed transcriptional arresthave been identified and characterized to some extent(18,30-33), the relative role of this mechanism in vivo indetermining the tissue-specific pattern of ADA expression is notclear. The search for the regulatory regions in the human ADAgene responsible for activation of transcription initiation hasrevealed two, a proximal region upstream of the transcriptionalstart site which has promoter function and a region within thefirst intron with T-cell specific enhancer function. Initialcharacterization of the enhancer domain has been previouslydescribed (4,9). Examination of the promoter for potentiallyfunctional elements by sequence homology revealed six putativeSpl binding sites and a poor homology TBP binding she (10,12).Somewhat similar elements have been identified in the mouseADA gene promoter (21). An equivalent to the ADA first intron



enhancer has not been identified in mouse. However, distalsequences upstream of the mouse ADA promoter have beenimplicated in regulation of ADA expression in some tissues intransgenic mouse studies (34).

Role of potential TBP binding siteInitial examination of the human ADA promoter revealed asequence TAAGAA, positioned at - 2 1 to — 26 bp, in theexpected location for a TATA box. At first glance this sequenceseems a plausible one to bind TBP, yet a weight matrixcomparison of this sequence and the surrounding sequence to theTBP binding region of other eukaryotic promoters (22) indicatedthat it is a poor consensus match. Mutation of this sequence toone with no homology to the TATA box sequence, CTGCAG,resulted in only a slight decrease in promoter activation. Thisindicates that this particular sequence is not essential for promoterfunction and probably does not bind TBP strongly, although itmay be loosely associated with it. Similar results have beenreported for other genes including SV40 early genes (35) andXenopus histone H2A (36). Deletion of the TBP binding site inthese genes does not affect in vivo levels of expression. However,in both, the absence of this site allows for the increased usageof several of the minor start sites (36—38). It is possible thatmutation of this sequence in the human ADA promoter has similarresults. We were unable to map the transcription start sites forthese TBP mutant clones directly to verify this hypothesis. Bycontrast, a similarly positioned sequence from the mouse ADAgene, TAAAAAA, is a better match to the TATA box consensussequence. It has been shown to bind TFI1D and to be requiredfor transcription of the mouse ADA gene (21). This result isinteresting in that it suggests that the murine ADA promoter andthe human ADA promoter may function by slightly differentmechanisms. This may in turn relate to some of the subtlevariations in tissue expression of ADA between mice and humans.

The function of the TATA box in promoters which containone is well studied. It is through TBP and some of its associatedfactors in the ThllD fraction commonly used for in vitro studiesthat both basal and activator-dependent transcription occur(39—43). The function of TBP in TATA-less promoters, of whichthe human ADA gene is an oft-cited example, is less wellunderstood. TBP, as part of the TFHD complex, has been provento be just as necessary for the function of these TATA-lesspromoters as for those with a TATA box (44). TBP is not thelimiting factor in transcription from these promoters, implyingthat another protein(s) that interact with TBP are limiting (40).These other proteins, termed TBP associated factors or TAFs,are also necessary for transcription from TATA-less promoters.How TBP is directed to a TATA-less promoter is currently understudy in many laboratories. Many divergent sequences have beenshown to bind TBP with lower affinity (45,46). In some casesbinding may represent a functional interaction between TBP andDNA in the absence of a specific recognition sequence (47) andchanges in the binding site sequence may have only minor effectson this function. If such an array of divergent sequences arecapable of recognizing and binding TBP, the question then ariseswhat constitutes a TBP binding site? A recently proposed model,suggests that the context of the site in vivo may affect the bindingof TBP to any given sequence (46). Binding of TBP to a weakbinding site may be improved by the addition of favorableupstream or downstream sequences such as a site that binds aninitiator protein or sites that bind activator proteins which areable to interact with TAFs. This model works admirably for the

human ADA promoter which seems to possess at best a veryweak TBP binding site. Mutation of this sequence did notsignificantly affect promoter activity, indicating that this sequencemay have little direct involvement in recruiting TBP to this site.This function is performed by surrounding sequences such as theSpl binding sites or possibly sequences at the transcriptional startsite which were not investigated. Since these sequences remainedthe same, TBP recruitment occured as usual and little loss ofactivation was observed.

Spl in basal promoter activationThe transcription factor Spl is required in both basal andenhanced activation of the ADA promoter. In the absence ofbinding sites for Spl, no basal activation of the promoter isobserved and enhancer-driven activation of the promoter isseriously compromised. All six of the Spl sites in the humanADA promoter have the ability to bind recombinant Spl. Analysisby deletion/truncation of these Spl sites indicates that the distalsites VI and V are not necessary for basal transcription and thefootprinting data suggests these sites may not bind Spl in vivo.Spl binds to the proximal sites I - IV synergistically to activatetranscription much as described for other promoters both natural(24,48-54) and artificial (44,55-57). In general more proximalsites play a more important role than those distal to the start siteand increasing the affinity of the most proximal site increasestranscriptional activation (50). This result is not surprising sinceSpl has been shown to interact with TAF's (42-44,55,58). Theability of Spl to do this is probably the reason the human ADAgene can exist without an effective TBP binding site and yetmaintain transcription. Basal activation of ADA should be highlydependent on the physiological concentration of Spl present ineach tissue type. Low concentrations of Spl would result in theoccupation of fewer sites and decreased synergistic activation ofthe ADA promoter. Spl protein and mRNA concentration inmouse tissues is highly variable (59). Some of the highest Splexpressing tissues in mouse, including columnar epithelial cellsof the upper gastrointestinal tract, maternal placenta, and thymus,are also among the highest ADA expressing tissues in mouse(60-63). Most of the mouse tissues which express high levelsof ADA also express high levels of Spl. However, there are alsotissues which express significant levels of Spl which express onlylow levels of ADA. Among these are lung and liver. Cells inthese tissues may maintain lower levels of ADA transcriptionunder high Spl concentration by utilization of other methods oftranscriptional regulation, perhaps transcriptional arrest orpausing (7,8,29,33). Therefore, as we have found in transientasssay, Spl is necessary but not sufficient for high level ADAexpression in vivo.

Promoter-enhancer interactionsThe interactions between enhancer bound proteins and promoterbound ones are just beginning to be studied. Many roles forproximally bound Spl have been suggested and one or more ofthese may explain the enhancer requirement for such Spl. Theinteraction of Spl with the TAF's may stabilize a component(s)of the transcriptional machinery (55). It is likely that theSpl — TAF interaction is responsible for the positioning of TBPnear the major transcription start site of the human ADA gene.Spl -TAF interactions at each of the multiple Spl sites mightalso explain the multiple transcriptional start sites observed inthis gene. This stabilization of the transcriptional machinery mightthen allow enhancer interactions with the assembled machinery



to occur more readily through different protein—proteininteraction(s). There is evidence to support this idea. We haveshown that the enhancer is able to activate transcription to lowlevels without proximal Spl sites present. This strongly suggeststhat at least some of the interactions of enhancer-bound proteinswith the transcriptional machinery in the promoter region occursindependently of bound Spl.

The favored model for enhancer-promoter interaction involvesthe association of bound factors at each site resulting in thelooping out of the intervening DNA. This ability has recentlybeen shown by Mastrangelo (64) and Su (65) for Spl, wherebydistantly bound Spl proteins interact to form multimers and loopout the intervening DNA. There is no doubt that the ADApromoter contains bound Spl, but the enhancer in its smallestdefined form does not contain a consensus Spl binding site (9).Sequences in the enhancer protected by methylation interference(ADA NF2) (9) are very similar but not identical to those reportedby Kingsley (66) for another member of the Spl family, Sp2.The mechanism for ADA enhancer function may involve Spl,or other members of its family, bound at or near the enhancerregion interacting with bound Spl in the proximal promoter, butthere is no direct evidence for this at the present.

It is unknown whether the enhancer requirement for Spl isspecific or if it reflects a generic requirement for an activator.All other promoters tested previously in heterologousconstructions with the CAT reporter gene and the ADA enhancercontained Spl sites in the promoter region (9). Analysis ofenhancer activation of a promoter(s) lacking Spl sites butcontaining the binding sequence for other transcriptional activatorsshould allow us to distinguish between the possibilities ofenhancer interaction with only basal machinery or with both themachinery and Spl. Studies at a detailed level await theidentification, purification, and characterization of the relevantenhancer-binding proteins.

This initial characterization of the human ADA promoter givesus insight into how the promoter may be activated to low andmoderate levels just by varying the levels of Spl in a tissue-specific manner. Preliminary studies of promoter—enhancerinteractions have given us a glimpse of how other regulatoryregions of the gene may interact with this basic unit to give highlevel activation in some tissues. More detailed characterizationof the elements in the ADA enhancer and their relationship toeach other will allow us to further examine the promoter-enhancer interaction.

ACKNOWLEDGEMENTS

This work was supported by National Institutes of Health grantGM 42969. Special thanks to R.Tjian for his generous gift ofthe recombinant Spl cDNA plasmid, pSpl-516C and to G.Schutzfor the plasmids pBLCATS and pBLCAT6 both of which wereinstrumental in these studies. Thanks to Amy Ruschulte andJennifer Ruschulte for their invaluable assistance in preparing,growing, and sequencing plasmids.

REFERENCES

1. Adams,A. and Harkness.R.A. (1976) din. Exp. immunoL, 26, 647-649.2. Van der Weyden.M.B. and Kelley.W.N. (1976) / Biol. Chan., 251,

5448-5456.3. Hirschhom.R., Martiniuk,F. and Rosen,F.S. (1978) din. ImmunoL

Immunopath., 9, 287-292.

4. ATODOW.B., Lattier.D., Silbiger.R., Dusing,M., Hutton.J., Jones.G.,StockJ., McNeislU., Potter.S., Witte.D. and Wiginton.D. (1989) GenesDev., 3, 1384-1400.

5. Giblett.E.R., Anderson J.E., Cohen,F., Pollara.B. and Meuwissen H J(1972) Lancet, 2, 1067-1069.

6. Martin.D.W. and Gelfand,E.W. (1981) Ann. Rev. Biochem., SO, 845-877.7. Lattier.D.L., States.J.C, HuttonJ.J. and Wiginton, D.A. (1989) Nucleic

Adds Res., 17, 1061-1076.8. Chen,Z., Harless,M.L., Wright.D.A. and Kellems.R.E. (1990) Mol. Cell.

Biol., 10, 4555-4564.9. Aronow.B.J., Silbiger,R.N., Dusing.M.R., Stock^.L., Yager.K.L.,

Potter.S.S., HuttonJ.J. and Wiginton.D.A. (1992) Mol. Cell Biol., 12,4170-4185.

10. Valerio.D., Duyvesteyn.M.G.C, Dekker.B.M.M., Weeda.G.,Berkvens.T.M., Van der Voorn,L., VanOrmondt.H. and Van der Eb.A.J.(1985) EMBO J., 4, 437-443.

11. Reitman,M.,Lee,E.,Westphal,H. and Felsenfeld,G. (1993) Mol Cell. Biol.,13, 3990-3998.

12. Wiginton.D.A., Kaplan,D.J., States.J.C, Akeson.A.L., Perme.C.M.,Bilyk.I.J., Vaughn,A.J., Lattier,D.L. and HuttonJ.J. (1986) Biochem., 25,8234-8244.

13. Dynan, W. S. (1986) Trends in Genet., 2, 196—197.14. Brcathnach,R. and Chambon.P. (1981) Annu. Rev. Biochem., SO, 349-383.15. Boshart.M., Kluppel.M., Schmidt.A., Schutz,G. and Luckow.B. (1992)

Gene, 110, 129-130.16. Kadonaga,J.T., Camer.K.R., Masiarz.F.R. and Tjian.R. (1987) Cell, 51,

1079-1090.17. Ingolk,D.E., Al-Ubaidi,M.R., Yeung.C, Bigo.H.A., Wright.D.A. and

KeUems.R.E. (1986) Mol. Cell. Biol., 6, 4458-4466.18. Maa,M.-C., ChinskyJ.M., Ramamurthy.V., Martin.B.D. and Kellems.R.E.

(1990)/ Biol. Chem., 265, 12513-12519.19. Ackerman.S.L., Minden.A.G., Williams.G.T., Bobonis.C. and Yeung.C.

(1991) Proc. Nail. Acad. Sd. USA, 88, 7523-7527.20. Al-Ubaidi,M.R., Ramamurthy,V., Maa.M., Ingolia.D.E., Chinsky.J.M.,

Martin.B.D. and Kellems.R.E. (1990) Genomics, 7, 476-485.21. InnisJ.W., Moore,D.J., Kash.S.F., Ramamurthy.V., Sawadogo.M. and

Kellems.R.E. (1991) J. Biol. Chem,, 266, 21765-21772.22. Bucher.P. (1990) J. Mol. Biol., 212, 563-578.23. Ciudad.C.J., Urlaub.G. and Chasin.L.A. (1988) J. Biol. Chem., 263,

16274-16282.24. Blake,M.C, Jambou.R.C, Swick.A.G., Kahn.J.W. and Azizkhan.J.C.

(1990) Mol. Cell. Biol., 10, 6632-6641.25. Briggs.M.R., KadonagaJX, Bell,S.P. and Tjian,R. (1986) Science, 234,

47-52.26. Kadonaga.J.T., Jones.K.A. and Tjian.R. (1986) Trends Biochem. Sd., 11,

20-23.27. Kriwacld.R.W., Schultz.S.C, Stehz.T.A. and CaradonnaJ.P. (1992) Proc.

Nail. Acad. Sd. USA, 89, 9759-9763.28. Gorman.C, Moffat.L.F. and Howard.B.H. (1982) Mol Cell. Biol, 2,

1044-105129. ChinskyJ.M., Maa,M.-C., Ramamurthy.V. and Kelkms,R.E. (1989) J. Biol

Chem., 264, 14561-14565.30. Chen.Z., innisJ.W., Sun.M., Wright.D.A. and Kellems.R.E. (1991) Mol

Cell. Biol, 11, 6248-6256.31. Ramamurthy.V., Maa,M.-C., Harless.M.L, Wright.D.A. and Kellems.R.E.

(1990) Mol. Cell. Biol, 10, 1484-1491.32. InrusJ.W. and Kellems.R.E. (1991) Mol Cell. Biol, 11, 5398-5409.33. Kash.S F., InnisJ.W., Jackson.A.U. and Kellems.R.E. (1993) Mol. Cell

Biol, 13, 2718-2729.34. WinstonJ.H., Hanten.G.R., Oerbeek.P.A. and KeIlems,R.E. (1992)/ Biol

Chem., 267, 13472-13479.35. Benoist.C. and Chambon.P. (1980) Proc. Noll Acad. Sd. USA, 77,

3865-3869.36. Grosschedl.R. and Bimstiel.M.L. (1980) Proc. Nail. Acad. Sd. USA, 77,

1432-1436.37. Benoist.C. and Chambon.P. (1981) Nature, 290, 304-310.38. Mathis.D.J. and Chambon.P. (1981) Nature, 290, 310-315.39. Buratowski.S., Hahn.S., Guarente.L. and Sharp.P.A. (1989) Cell, 56,

549-561.40. ColgaM. and Manley^.L. (1992) Genes Dev., 6, 304-315.41. Sawadogo.M. and Roeder.R.G. (1985) Cell, 43, 165-175.42. GUI.G. and Tjian.R. (1992) Current Opinions in Genetics and Dev., 2,

236-242.43. Dynlaeht,B.D., Hoey.T. and Tjian,R. (1991) Cell, 66, 563-576.44. Pugh.B.F. and Tjian.R. (1991) Genes Dev., 5, 1935-1945.



45. Hahn.S., Buratowski.S., Sharp,P.A. and Guarente.L. (1989) Proc. Natl.Acad. Sci. USA, 86, 5718-5722.

46. Wiley,S.R., Kraus.R.J. and Mertz,J.E. (1992) Proc. NatL Acad. Sci. USA,89, 5814-5818.

47. Zenzie-Grcgory.B., Khachi,A., GanawayJ.P. and Smale.S.T. (1993) Mol.Cell. Biol, 13, 3841-3849.

48. Tasanen,K., OikarinenJ., Kivirikko.K.I. and Pihlajaniemi.T. (1993) Biochem.J., 292, 41-45.

49. Pogulis.R.J. and Freytag.S.O. (1993) J. Biol. Oiem., 268, 2493-2499.50. Gidoni.D., KadonagaJ.T., Bancra-Saldana.H., Takahashi.K., Chambon.P.

and Tjian.R. (1985) Science, 230, 511-517.51. Proudfoot.N.J., Lee.B.A. and Monks.J. (1992) New Biol., 4, 369-381.52. Chen.X., Azizkhan.J.C. and Lee.D.C. (1992) Oncogene, 7, 1805-1815.53. Jones.K.A., Yamamoto.K.R. and Tjian.R. (1985) Cell, 42, 559-572.54. Anderson.G.M. and Frcytag.S.O. (1991) Mol. Cell Biol, 11, 1935-1943.55. Pugh.B.F. and Tjian,R. (1990) Cell, 61, 1187-1197.56. O'Shea-Greenfield.A. and Smale.S.T. (1992) /. Biol. Oiem., 267,

1391-1402.57. Smale,S.T. and Baltimore.D. (1989) Cell, 57, 103-113.58. Pugh.B.F. and Tjian.R. (1992) / Biol. Oiem., 267, 679-682.59. SafferJ.D., Jackson.S.P. and Annarella.M.B. (1991) Mol. Cell. Biol, 11,

2189-2199.60. Brady,T.G. and O'Donovan,C.I. (1965) Comp. Biochem. Physiol., 14,

101-120.61. Knudsen.T.B., GreenJ.D., Airhart.M.J., Higley.H.R., ChinskyJ.M. and

Kdlems.R.E. (1988) Biol. ofReprod., 39, 937-951.62. Chinsky.J.M., Ramamurthy.V., Fanslow.W.C, Ingolia.D.E.,

Blackbum.M.R., Shaffer.K.T., Higley.H.R., TrentinJ.J., Rudolph.F.B.,Knudsen.T.B. and Kellems.R.E. (1990) Differentiation, 42, 172-183.

63. Witte.D.P., Wiginton.D.A., Hunon.J.J. and Axonow.BJ. (1991) / Cell.Biol., 115, 179-190.

64. Mastrangelo,I.A., Courey.A.J., Wall J.S., Jacksoo,S.P. and Hough.P.V.C.(1991) Proc. Nasl. Acad. Sci. USA, 88, 5670-5674.

65. Su,W., Jackson.S., Tjian,R. and Echols.H. (1991) GenesDev., 5, 820-826.66. Kingsley.C. and Winoto,A. (1992) Mol. Cell. Biol, 12, 4251-4261.


Spi is essential for both enhancer-mediated and basal activation of ...

Documents

Transcript of Spi is essential for both enhancer-mediated and basal activation of ...